Medical image fusion based on quaternion wavelet transform. (June 2020)

• Record Type: Journal Article
• Authors: Zhancheng, Zhang; Xiaoqing, Luo; Mengyu, Xiong; Zhiwen, Wang; Kai, Li
• Is Part Of: Journal of algorithms & computational technology. Volume 14(2020)
• Publication Date: 2020-06
• Subjects: Medical image fusion -- quaternion wavelet transform -- context -- activity measure; Computer algorithms -- Periodicals; Numerical calculations -- Periodicals; Computer algorithms; Numerical calculations; Periodicals; 518.1
• DOI: 10.1177/1748302620931297
• Languages: English
• ISSNs: 1748-3018

Abstract

Medical image fusion can combine multi-modal images into an integrated higher-quality image, which can provide more comprehensive and accurate pathological information than individual image does. Traditional transform domain-based image fusion methods usually ignore the dependencies between coefficients and may lead to the inaccurate representation of source image. To improve the quality of fused image, a medical image fusion method based on the dependencies of quaternion wavelet transform coefficients is proposed. First, the source images are decomposed into low-frequency component and high-frequency component by quaternion wavelet transform. Then, a clarity evaluation index based on quaternion wavelet transform amplitude and phase is constructed and a contextual activity measure is designed. These measures are utilized to fuse the high-frequency coefficients and the choose-max fusion rule is applied to the low-frequency components. Finally, the fused image can be obtained by inverse quaternion wavelet transform. The experimental results on some brain multi-modal medical images demonstrate that the proposed method has achieved advanced fusion result.